Skip to content

Conversation

@hansenp
Copy link

@hansenp hansenp commented Nov 25, 2025

I implemented the following method to re-scale the BOQA scores so that they can be combined with the variant score in Exomiser.

image

The boqaRawLogScores are calculated directly in BOQA from the four counts to avoid numerical instabilities.

This score has the following properties:

  1. Exomiser compatible: The maximum score is always 1.0, the lowest is 0.
  2. BOQA's ranking remains unchanged.
  3. The original BOQA scores are somehow reflected in the transformed BoqaExomiserScore (which is not the case with rank scaled score).

Here is the distribution of scores across ranks for 3000 phenopackets from phenopacket-score for which BOQA ranks the diagnosed disease first.

image

To implement this, I added a function in BOQA BoqaPatientAnalyzer that provides the BoqaPrioritizer in Exomiser with unnormalized BOQA log scores. See here: https://github.com/P2GX/BOQA/pull/32

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants